Ermi1223/amharic-hate-speech-api

# Amharic Hate Speech Detection API This project is an **Amharic language-based hate speech detection** system, designed to identify and classify offensive or harmful language in Amharic text. Built using the **Flask** framework and leveraging **BERT** models, the API detects hate speech in user-submitted text and categorizes it as either **"ጥላቻ" (Hate Speech)** or **"መልካም" (Non-Hate Speech)**. The API integrates seamlessly into applications where text moderation or automated content review is necessary, especially for platforms serving Ethiopian and Amharic-speaking communities. ## Key Features: - **Real-time Hate Speech Detection**: Classifies Amharic text into hate speech and non-hate speech categories. - **Flask-Based API**: A lightweight, easy-to-use API built with Flask for integration into various applications. - **BERT Model**: Utilizes pre-trained BERT models fine-tuned on an Amharic hate speech dataset for accurate text classification. - **Simple Interface**: Easy-to-use endpoints that accept text input and return predictions. ## Installation and Usage: 1. Clone the repository: ```bash git clone github.com ``` 2. Install dependencies: ```bash pip install -r requirements.txt ``` 3. Run the Flask app: ```bash python app.py ``` 4. Send POST requests to the `/predict` endpoint with Amharic text to receive predictions. ## Technologies: - **Flask**: Lightweight web framework for building the API. - **Transformers (Hugging Face)**: Used for loading the pre-trained BERT model and tokenizer. - **PyTorch**: Backend framework for the model inference. - **Streamlit** (if integrating with a front-end): For building the UI to interact with the API. ## Example Request: ```bash POST localhost Content-Type: application/json { "text": "ሰላም ሰዎች እንዴት ነህ/ነሽ?" } ``` ## Response: ```json { "prediction": "መልካም" } ```